Search

In conjunction with pandemics, medical image data are growing exponentially. In some countries, hospitals collect biometric data from patients, such as fingerprints, iris, or faces. This data can be used for things like identity verification and security management. However, this medical data can be easily compromised by hackers. In order to prevent illegal tampering with medical images and invasion of privacy, a new texture fusion medical image encryption (TFMIE) algorithm derived from biometric images is proposed, which can encrypt the image using biometric information for storage or transmission. First, the medical image is decomposed into n-bit-planes by bit-plane decomposition. Secondly, a fusion image is generated by a biometric image with a circular local binary pattern and pixel-weighted average method. The fused image is further decomposed into n bit-planes through bit-plane decomposition and performs XOR operation with the original medical image in reverse order. Following the execution of the XOR operation, a new scrambling and diffusion algorithm based on a one-dimensional fractional trigonometric function (1DFTF) chaotic map is employed to form the cipher image. The experimental results show that compared with the existing methods, the average information entropy value of TFMIE is 7.99, and the average values of NPCR and UACI reach 0.9958 and 0.3346, respectively, which have strong key sensitivity, good robustness, and anti-attack ability. The method is lossless and has high transmission efficiency, which can meet the needs of medical big data encryption. [ABSTRACT FROM AUTHOR]

Objective: To study the clinical application value of enhanced computed tomography (CT) image texture parameters combined with serum carcinoembryonic antigen (CEA) and carbohydrate antigen 724 (CA724) in the recurrence and metastasis of esophageal cancer after operation. Methods: 86 patients with esophageal cancer who underwent operation treatment in our hospital from June 2019 to July 2021 were selected. The serum CEA and CA724 levels were detected before operation, and enhanced CT scanning was performed, the image data of the arteriovenous phase were analyzed by texture, and the mean value, median, standard deviation, unevenness, skewness, kurtosis, entropy and other parameters were recorded. The patients with esophageal cancer were followed up for 12 months after operation. According to whether they had recurrence and metastasis, they were divided into recurrence and metastasis group (n=32) and non-recurrence and metastasis group (n=54). The differences of texture parameters of enhanced CT images and serum CEA and CA724 levels in the two groups were compared. Receiver Operating characteristic (ROC) curve was used to analyze the predictive value of enhanced CT image texture parameters combined with serum CEA and CA724 for postoperative recurrence and metastasis of patients with esophageal cancer. Results: The arterial stage entropy in the recurrence and metastasis group was higher than that in the non-recurrence and metastasis group (P<0.05). The inevenness and entropy of venous phase in the recurrence and metastasis group were higher than those in the non-recurrence and metastasis group(P<0.05). The serum CEA and CA724 levels in the recurrence and metastasis group were higher than those in the non-recurrence and metastasis group (P<0.05). ROC curve showed that the combination of serum CEA and CA724 had significantly improved the arterial phase entropy, the venous phase unevenness, the area under the curve (AUC) of the entropy, the sensitivity and the specificity compared with the single application of each index, and the combined application had higher predictive efficiency. Conclusion: The serum CEA and CA724 levels and the increase of the entropy and the inhomogeneity and the arterial phase entropy and the inhomogeneity and the entropy of venous phase are related to the recurrence and metastasis of the postoperative patients with esophageal cancer. The combination of the arterial phase entropy and the inhomogeneity and the entropy of venous phase with the serum CEA and CA724 has of high application value in predicting the recurrence and metastasis of the postoperative patients with esophageal cancer. [ABSTRACT FROM AUTHOR]

Objectives: To compare deep learning (True Fidelity, TF) and partial model based Iterative Reconstruction (ASiR-V) algorithm for image texture, low contrast lesion detectability and potential dose reduction.Methods: Anthropomorphic phantoms (mimicking non-overweight and overweight patient), containing lesions of 6 mm in diameter with 20HU contrast, were scanned at five different dose levels (2,6,10,15,20 mGy) on a CT system, using clinical routine protocols for liver lesion detection. Images were reconstructed using ASiR-V 0% (surrogate for FBP), 60 % and TF at low, medium and high strength. Noise texture was characterized by computing a normalized Noise Power Spectrum filtered by an eye filter. The similarity against FBP texture was evaluated using peak frequency difference (PFD) and root mean square deviation (RMSD). Low contrast detectability was assessed using a channelized Hotelling observer and the area under the ROC curve (AUC) was used as figure of merit. Potential dose reduction was calculated to obtain the same AUC for TF and ASiR-V.Results: FBP-like noise texture was more preserved with TF (PFD from -0.043mm-1 to -0.09mm-1, RMSD from 0.12mm-1 to 0.21mm-1) than with ASiR-V (PFD equal to 0.12 mm-1, RMSD equal to 0.53mm-1), resulting in a sharper image. AUC was always higher with TF than ASIR-V. In average, TF compared to ASiR-V, enabled a radiation dose reduction potential of 7%, 25 % and 33 % for low, medium and high strength respectively.Conclusion: Compared to ASIR-V, TF at high strength does not impact noise texture and maintains low contrast liver lesions detectability at significant lower dose. [ABSTRACT FROM AUTHOR]

Introduction: Chest radiograph and computed tomography (CT) scans can accidentally reveal pulmonary nodules. Malignant and benign pulmonary nodules can be difficult to distinguish without specific imaging features, such as calcification, necrosis, and contrast enhancement. However, these lesions may exhibit different image texture characteristics which cannot be assessed visually. Thus, a computer‐assisted quantitative method like histogram analysis (HA) of Hounsfield unit (HU) values can improve diagnostic accuracy, reducing the need for invasive biopsy. Methods: In this exploratory control study, nonenhanced chest CT images of 20 patients with benign (10) and cancerous (10) lesion were selected retrospectively. The appearances of benign and malignant lesions were very similar in chest CT images, and only pathology report was used to discriminate them. Free hand region of interest (ROI) was inserted inside the lesion for all slices of each lesion. Mean, minimum, maximum, and standard deviations of HU values were recorded and used to make HA. Results: HA showed that the most malignant lesions have a mean HU value between 30 and 50, a maximum HU less than 150, and a minimum HU between −30 and 20. Lesions outside these ranges were mostly benign. Conclusion: Quantitative CT analysis may differentiate malignant from benign lesions without specific malignancy patterns on unenhanced chest CT image. [ABSTRACT FROM AUTHOR]

Objective: To evaluate image quality and diagnostic confidence improvement using a thin slice and a deep learning image reconstruction (DLIR) in contrast-enhanced abdominal CT. Methods: Forty patients with hepatic lesions in enhanced abdominal CT were retrospectively analyzed. Images in the portal phase were reconstructed at 5 mm and 1.25 mm slice thickness using the 50% adaptive statistical iterative reconstruction (ASIR-V) (ASIR-V50%) and at 1.25 mm using DLIR at medium (DLIR-M) and high (DLIR-H) settings. CT number and standard deviation of the hepatic parenchyma, spleen, portal vein, and subcutaneous fat were measured, and contrast-to-noise ratio (CNR) was calculated. Edge-rise-slope (ERS) was measured on the portal vein to reflect spatial resolution and the CT number skewness on liver parenchyma was calculated to reflect image texture. Two radiologists blindly assessed the overall image quality including subjective noise, image contrast, visibility of small structures using a 5-point scale, and object sharpness and lesion contour using a 4-point scale. Results: For the 1.25-mm images, DLIR significantly reduced image noise, improved CNR and overall subjective image quality compared to ASIR-V50%. Compared to the 5-mm ASIR-V50% images, DLIR images had significantly higher scores in the visibility and contour for small structures and lesions; as well as significantly higher ERS and lower CT number skewness. At a quarter of the signal strength, the 1.25-mm DLIR-H images had a similar subjective noise score as the 5-mm ASIR-V50% images. Conclusion: DLIR significantly reduces image noise and maintains a more natural image texture; image spatial resolution and diagnostic confidence can be improved using thin slice images and DLIR in abdominal CT. Key Points: • DLIR further reduces image noise compared with ASIR-V while maintaining favorable image texture. • In abdominal CT, thinner slice images improve image spatial resolution and small object visualization but suffer from higher image noise. • Thinner slice images combined with DLIR in abdominal CT significantly suppress image noise for detecting low-density lesions while significantly improving image spatial resolution and overall image quality. [ABSTRACT FROM AUTHOR]

Objectives: To perform a comprehensive within-subject image quality analysis of abdominal CT examinations reconstructed with DLIR and to evaluate diagnostic accuracy compared to the routinely applied adaptive statistical iterative reconstruction (ASiR-V) algorithm. Materials and methods: Oncologic patients were prospectively enrolled and underwent contrast-enhanced CT. Images were reconstructed with DLIR with three intensity levels of reconstruction (high, medium, and low) and ASiR-V at strength levels from 10 to 100% with a 10% interval. Three radiologists characterized the lesions and two readers assessed diagnostic accuracy and calculated signal-to-noise ratio (SNR), contrast-to-noise ratio (CNR), figure of merit (FOM), and subjective image quality, the latter with a 5-point Likert scale. Results: Fifty patients (mean age: 70 ± 10 years, 23 men) were enrolled and 130 liver lesions (105 benign lesions, 25 metastases) were identified. DLIR_H achieved the highest SNR and CNR, comparable to ASiR-V 100% (p ≥.051). DLIR_M returned the highest subjective image quality (score: 5; IQR: 4–5; p ≤.001) and significant median increase (29%) in FOM (p <.001). Differences in detection were identified only for lesions ≤ 0.5 cm: 32/33 lesions were detected with DLIR_M and 26 lesions were detected with ASiR-V 50% (p =.031). Lesion accuracy of was 93.8% (95% CI: 88.1, 97.3; 122 of 130 lesions) for DLIR and 87.7% (95% CI: 80.8, 92.8; 114 of 130 lesions) for ASiR-V 50%. Conclusions: DLIR yields superior image quality and provides higher diagnostic accuracy compared to ASiR-V in the assessment of hypovascular liver lesions, in particular for lesions ≤ 0.5 cm. Clinical relevance statement: Deep learning image reconstruction algorithm demonstrates higher diagnostic accuracy compared to iterative reconstruction in the identification of hypovascular liver lesions, especially for lesions ≤ 0.5 cm. Key Points: • Iterative reconstruction algorithm impacts image texture, with negative effects on diagnostic capabilities. • Medium-strength deep learning image reconstruction algorithm outperforms iterative reconstruction in the diagnostic accuracy of ≤ 0.5 cm hypovascular liver lesions (93.9% vs 78.8%), also granting higher objective and subjective image quality. • Deep learning image reconstruction algorithm can be safely implemented in routine abdominal CT protocols in place of iterative reconstruction. [ABSTRACT FROM AUTHOR]

Background: Radiation dose reduction has been the focus of many research activities in x‐ray CT. Various approaches were taken to minimize the dose to patients, ranging from the optimization of clinical protocols, refinement of the scanner hardware design, and development of advanced reconstruction algorithms. Although significant progress has been made, more advancements in this area are needed to minimize the radiation risks to patients. Purpose: Reconstruction algorithm‐based dose reduction approaches focus mainly on the suppression of noise in the reconstructed images while preserving detailed anatomical structures. Such an approach effectively produces synthesized high‐dose images (SHD) from the data acquired with low‐dose scans. A representative example is the model‐based iterative reconstruction (MBIR). Despite its widespread deployment, its full adoption in a clinical environment is often limited by an undesirable image texture. Recent studies have shown that deep learning image reconstruction (DLIR) can overcome this shortcoming. However, the limited availability of high‐quality clinical images for training and validation is often the bottleneck for its development. In this paper, we propose a novel approach to generate SHD with existing low‐dose clinical datasets that overcomes both the noise texture issue and the data availability issue. Methods: Our approach is based on the observation that noise in the image can be effectively reduced by performing image processing orthogonal to the imaging plane. This process essentially creates an equivalent thick‐slice image (TSI), and the characteristics of TSI depend on the nature of the image processing. An advantage of this approach is its potential to reduce impact on the noise texture. The resulting image, however, is likely corrupted by the anatomical structural degradation due to partial volume effects. Careful examination has shown that the differential signal between the original and the processed image contains sufficient information to identify regions where anatomical structures are modified. The differential signal, unfortunately, contains significant noise and has to be removed. The noise removal can be accomplished by performing iterative noise reduction to preserve structural information. The processed differential signal is subsequently subtracted from TSI to arrive at SHD. Results: The algorithm was evaluated extensively with phantom and clinical datasets. For better visual inspection, difference images between the original and SHD were generated and carefully examined. Negligible residual structure could be observed. In addition to the qualitative inspection, quantitative analyses were performed on clinical images in terms of the CT number consistency and the noise reduction characteristics. Results indicate that no CT number bias is introduced by the proposed algorithm. In addition, noise reduction capability is consistent across different patient anatomical regions. Further, simulated water phantom scans were utilized in the generation of the noise power spectrum (NPS) to demonstrate the preservation of the noise‐texture. Conclusions: We present a method to generate SHD datasets from regularly acquired low‐dose CT scans. Images produced with the proposed approach exhibit excellent noise‐reduction with the desired noise‐texture. Extensive clinical and phantom studies have demonstrated the efficacy and robustness of our approach. Potential limitations of the current implementation are discussed and further research topics are outlined. [ABSTRACT FROM AUTHOR]

We propose an artificial intelligence prediction method for extracorporeal shock wave lithotripsy treatment outcomes by analysis of a wide variety of variables. We retrospectively reviewed the records of 171 patients from between January 2009 and November 2019 that underwent shock wave lithotripsy at Wakayama Medical University, Japan, for ureteral stones shown on preoperative non-contrast computed tomography. This prediction method consisted of stone area extraction, stone analyzing factor extraction from non-contrast computed tomography images, and shock wave lithotripsy treatment result prediction by a non-linear support vector machine for analysis of 15 input and automatic measurement factors. Input factors included patient age, skin-to-stone distance, and maximum ureteral wall thickness, and the automatic measurement factors included 11 non-contrast computed tomography image texture factors in the stone area and stone volume. Permutation feature importance was also applied to the artificial intelligence prediction results to analyze the importance of each factor relating to estimate decision grounds. The prediction performance was evaluated by five-fold cross-validation, it obtained 0.742 of the mean area under the receiver operating characteristic curve. The proposed method is shown by these results to have robust data diversity and effective clinical application. As a result of permutation feature importance, some factors that showed high p-values in the significant difference tests were thought to have a high contribution to the proposed prediction method. Future issues include validation using a larger volume of high-resolution clinical non-contrast computed tomography image data and the application of deep learning. [ABSTRACT FROM AUTHOR]

The objective of this study is to create patient-specific phantoms for computed tomography (CT) that possess accurate densities and exhibit visually realistic image textures. These qualities are crucial for evaluating CT performance in clinical settings. The study builds upon a previously presented 3D printing method (PixelPrint) by incorporating soft tissue and bone structures. We converted patient DICOM images directly into 3D printer instructions using PixelPrint and utilized calcium-doped filament to increase the Hounsfield unit (HU) range. Density was modeled by controlling printing speed according to volumetric filament ratio to emulate attenuation profiles. We designed micro-CT phantoms to demonstrate the reproducibility, and to determine mapping between filament ratios and HU values on clinical CT systems. Patient phantoms based on clinical cervical spine and knee examinations were manufactured and scanned with a clinical spectral CT scanner. The CT images of the patient-based phantom closely resembled original CT images in visual texture and contrast. Micro-CT analysis revealed minimal variations between prints, with an overall deviation of ± 0.8% in filament line spacing and ± 0.022 mm in line width. Measured differences between patient and phantom were less than 12 HU for soft tissue and 15 HU for bone marrow, and 514 HU for cortical bone. The calcium-doped filament accurately represented bony tissue structures across different X-ray energies in spectral CT (RMSE ranging from ± 3 to ± 28 HU, compared to 400 mg/ml hydroxyapatite). In conclusion, this study demonstrated the possibility of extending 3D-printed patient-based phantoms to soft tissue and bone structures while maintaining accurate organ geometry, image texture, and attenuation profiles. [ABSTRACT FROM AUTHOR]

Objective. While simulated low-dose CT images and phantom studies cannot fully approximate subjective and objective effects of deep learning (DL) denoising on image quality, live animal models may afford this assessment. This study is to investigate the potential of DL in CT dose reduction on image quality compared to iterative reconstruction (IR). Approach. The upper abdomen of a live 4 year old sheep was scanned on a CT scanner at different exposure levels. Images were reconstructed using FBP and ADMIRE with 5 strengths. A modularized DL network with 5 modules was used for image reconstruction via progressive denoising. Radiomic features were extracted from a region over the liver. Concordance correlation coefficient (CCC) was applied to quantify agreement between any two sets of radiomic features. Coefficient of variation was calculated to measure variation in a radiomic feature series. Structural similarity index (SSIM) was used to measure the similarity between any two images. Diagnostic quality, low-contrast detectability, and image texture were qualitatively evaluated by two radiologists. Pearson correlation coefficient was computed across all dose-reconstruction/denoising combinations. Results. A total of 66 image sets, with 405 radiomic features extracted from each, are analyzed. IR and DL can improve diagnostic quality and low-contrast detectability and similarly modulate image texture features. In terms of SSIM, DL has higher potential in preserving image structure. There is strong correlation between SSIM and radiologists’ evaluations for diagnostic quality (0.559) and low-contrast detectability (0.635) but moderate correlation for texture (0.313). There is moderate correlation between CCC of radiomic features and radiologists’ evaluation for diagnostic quality (0.397), low-contrast detectability (0.417), and texture (0.326), implying that improvement of image features may not relate to improvement of diagnostic quality. Conclusion. DL shows potential to further reduce radiation dose while preserving structural similarity, while IR is favored by radiologists and more predictably alters radiomic features. [ABSTRACT FROM AUTHOR]

The objective of this study is to create patient-specific phantoms for computed tomography (CT) that possess accurate densities and exhibit visually realistic image textures. These qualities are crucial for evaluating CT performance in clinical settings. The study builds upon a previously presented 3D printing method (PixelPrint) by incorporating soft tissue and bone structures. We converted patient DICOM images directly into 3D printer instructions using PixelPrint and utilized calcium-doped filament to increase the Hounsfield unit (HU) range. Density was modeled by controlling printing speed according to volumetric filament ratio to emulate attenuation profiles. We designed micro-CT phantoms to demonstrate the reproducibility, and to determine mapping between filament ratios and HU values on clinical CT systems. Patient phantoms based on clinical cervical spine and knee examinations were manufactured and scanned with a clinical spectral CT scanner. The CT images of the patient-based phantom closely resembled original CT images in visual texture and contrast. Micro-CT analysis revealed minimal variations between prints, with an overall deviation of ± 0.8% in filament line spacing and ± 0.022 mm in line width. Measured differences between patient and phantom were less than 12 HU for soft tissue and 15 HU for bone marrow, and 514 HU for cortical bone. The calcium-doped filament accurately represented bony tissue structures across different X-ray energies in spectral CT (RMSE ranging from ± 3 to ± 28 HU, compared to 400 mg/ml hydroxyapatite). In conclusion, this study demonstrated the possibility of extending 3D-printed patient-based phantoms to soft tissue and bone structures while maintaining accurate organ geometry, image texture, and attenuation profiles. [ABSTRACT FROM AUTHOR]

This study retrospectively included some patients with colorectal cancer diagnosed by histopathology, to explore the feasibility of CT medical image texture analysis in predicting KRAS gene mutations in patients with colorectal cancer. Before any surgical procedure, all patients received an enhanced CT scan of the abdomen and pelvis, as well as genetic testing. To define patient groups, divide all patients into test and validation sets based on the order of patient enrollment. A radiologist took a look at the plain axial CT image of the tumor, as well as the portal vein CT image, at the corresponding level. The physician points the computer's cursor to the relevant area in the image, and TexRAD software programs together texture parameters based on various spatial scale factors, also known as total mean, total variance, statistical entropy, overall total average, mean total, positive mean, skewness value, kurtosis value, and general skewness. Using the same method again two weeks later, the observer and another physician measured the image of each patient again to see if the method was consistent between observers. With regard to clinical information, the KRAS gene mutation group and the wild group of participants in the test set and validation set each had values for the texture parameter. In a study of patients with colorectal cancer, the results demonstrated that CT texture parameters were correlated with the presence of the KRAS gene mutation. The best CT prediction model includes the values of the medium texture image's slope and the other CT fine texture image's value of entropy, the medium texture image's slope and kurtosis, and the medium texture image's mean and the other CT fine texture image's value of entropy. Regardless of the training set or the validation set, patients with and without KRAS gene mutations did not differ significantly in clinical characteristics. This method can be used to identify mutations in the KRAS gene in patients with colorectal cancer, making it practical to implement CT medical image texture analysis technology for that purpose. [ABSTRACT FROM AUTHOR]

Background: Deep learning surpasses many traditional methods for many vision tasks, allowing the transformation of hierarchical features into more abstract, high-level features. Purpose: To evaluate the prognostic value of preoperative computed tomography (CT) image texture features and deep learning self-learning high-throughput features (SHF) on postoperative overall survival in the treatment of patients with colorectal cancer (CRC). Material and Methods: The dataset consisted of 810 enrolled patients with CRC confirmed from 10 November 2011 to 10 February 2018. In contrast, SHF extracted by deep learning with multi-task training mechanism and texture features were extracted from the CT with tumor volume region of interest, respectively, and combined with the Cox proportional hazard (CoxPH) model for initial validation to obtain a RAD score to classify patients into high- and low-risk groups. The SHF stability was further validated in combination with Neural Multi-Task Logistic Regression (N-MTLR) model. The overall recognition ability and accuracy of CoxPH and N-MTLR model were evaluated by C-index and Integrated Brier Score (IBS). Results: SHF had a more significant degree of differentiation than texture features. The result is (SHF vs. texture features: C-index: 0.884 vs. 0.611; IBS: 0.025 vs. 0.073) in the CoxPH model, and (SHF vs. texture features: C-index: 0.861 vs. 0.630; IBS: 0.024 vs. 0.065) in N-MTLR. Conclusion: SHF is superior to texture features and has potential application for the preoperative prediction of the individualized treatment of CRC. [ABSTRACT FROM AUTHOR]

Objectives: To evaluate image quality and diagnostic performance of carotid dual-energy computed tomography angiography (DECTA) using deep learning image reconstruction (DLIR) compared with images using adaptive statistical iterative reconstruction-Veo (ASIR-V). Methods: Carotid DECTA datasets of 28 consecutive patients were reconstructed at 50 keV using DLIR at low, medium, and high levels (DLIR-L, DLIR-M, and DLIR-H) and 80% ASIR-V algorithms. Mean attenuation, image noise, signal-to-noise ratio (SNR), and contrast-to-noise ratio (CNR) at different levels of arteries were measured and calculated. Image quality for noise and texture, depiction of arteries, and diagnostic performance toward carotid plaques were assessed subjectively by two radiologists. Quantitative and qualitative parameters were compared between the ASIR-V, DLIR-L, DLIR-M, and DLIR-H groups. Results: The image noise at aorta and common carotid artery, SNR, and CNR at all level arteries of DLIR-H images were significantly higher than those of ASIR-V images (p = 0.000–0.040). The quantitative analysis of DLIR-L and DLIR-M showed comparable denoise capability with ASIR-V. The overall image quality (p = 0.000) and image noise (p = 0.000–0.014) were significantly better in the DLIR-M and DLIR-H images. The image texture was improved by DLR at all level compared to ASIR-V images (p = 0.000–0.008). Depictions of head and neck arteries and diagnostic performance were comparable between four groups (p > 0.05). Conclusions: Compared with 80% ASIR-V, we recommend DLIR-H for clinical carotid DECTA reconstruction, which can significantly improve the image quality of carotid DECTA at 50 keV but maintain a desirable diagnostic performance and arterial depiction. [ABSTRACT FROM AUTHOR]

This study aimed to evaluate the value of the deep learning image reconstruction (DLIR) algorithm (GE Healthcare's TrueFidelity™) in improving the image quality of low-dose computed tomography (LDCT) of the chest. First, we retrospectively extracted raw data of chest LDCT from 50 patients and reconstructed them by using model-based adaptive statistical iterative reconstruction-Veo at 50% (ASIR-V 50%) and DLIR at medium and high strengths (DLIR-M and DLIR-H). Three sets of images were obtained. Next, two radiographers measured the mean CT value/image signal and standard deviation (SD) in Hounsfield units at the region of interest (ROI) and calculated the signal-to-noise ratio (SNR) and contrast-to-noise ratio (CNR). Two radiologists subjectively evaluated the image quality using a 5-point Likert scale. The differences between the groups of data were analyzed through a repeated measures ANOVA or the Friedman test. Last, our result show that the three reconstructions did not differ significantly in signal (p > 0.05) but had significant differences in noise, SNR, and CNR (p < 0.001). The subjective scores significantly differed among the three reconstruction modalities in soft tissue (p < 0.001) but not in lung tissue (p > 0.05). DLIR-H had the best noise reduction ability and improved SNR and CNR without distorting the image texture, followed by DLIR-M and ASIR-V 50%. In summary, DLIR can provide a higher image quality at the same dose, enhancing the physicians' diagnostic confidence and improving the diagnostic efficacy of LDCT for lung cancer screening. [ABSTRACT FROM AUTHOR]

Background: This study investigated the prognostic value of axillary lymph node (ALN) heterogeneity texture features through 18F-fluorodeoxyglucose positron emission tomography/computed tomography (18F-FDG PET/CT) in patients with locally advanced breast cancer (LABC). Methods: We retrospectively analyzed 158 LABC patients with FDG-avid, pathology-proven, metastatic ALN who underwent neoadjuvant chemotherapy (NAC) and curative surgery. Tumor and ALN texture parameters were extracted from pretreatment 18F-FDG PET/CT using Chang-Gung Image Texture Analysis software. The least absolute shrinkage and selection operator regression was performed to select the most significant predictive texture parameters. The predictive impact of texture parameters was evaluated for both progression-free survival and pathologic NAC response. Results: The median follow-up period of 36.8 months and progression of disease (PD) was observed in 36 patients. In the univariate analysis, ALN textures (minimum standardized uptake value (SUV) (p = 0.026), SUV skewness (p = 0.038), SUV bias-corrected Kurtosis (p = 0.034), total lesion glycolysis (p = 0.011)), tumor textures (low-intensity size zone emphasis (p = 0.045), minimum SUV (p = 0.047), and homogeneity (p = 0.041)) were significant texture predictors. On the Cox regression analysis, ALN SUV skewness was an independent texture predictor of PD (p = 0.016, hazard ratio 2.3, 95% confidence interval 1.16–4.58). Conclusions: ALN texture feature from pretreatment 18F-FDG PET/CT is useful for the prediction of LABC progression. [ABSTRACT FROM AUTHOR]

Objective: Evaluate and compare the image quality and acceptance of a full MBIR algorithm to that of an earlier full IR hybrid algorithm and filtered back projection (FBP).Methods: Acquisitions were performed with a 320 detector-row CT scanner with seven different dose levels. Images were reconstructed with three algorithms: FBP, full hybrid iterative reconstruction (HIR), and a full model-based iterative reconstruction algorithm (full MBIR). The sensitometry, spatial resolution, image texture, and low-contrast detectability of these algorithms were compared. Subjective analysis of low-contrast detectability was performed. Ten radiologists answered a questionnaire on image quality and confidence in full MBIR images in clinical practice.Results: The contrast-to-noise ratio of full MBIR was significantly higher than in the other algorithms (p < 0.0015). The spatial resolution was also higher with full MBIR at high frequencies (> 0.3 lp/mm). Full MBIR at low dose levels led to better low-contrast detectability and more inserts being identified with a higher confidence (p < 0.0001). Full MBIR was associated with a change in image texture compared to HIR and FBP. Eighty percent of radiologists judged general appearance and texture of full MBIR images worse than HIR. Moreover, compared with HIR, for 50% of radiologists, the diagnostic confidence on full MBIR images was worse. Questionnaire reliability was considered acceptable (Cronbach alpha 0.7).Conclusion: Compared to conventional iterative reconstruction algorithms, full MBMIR presented a higher image quality and low-contrast detectability and a worse acceptance among radiologists.Key Points: • Full MBIR used led to an overall improvement in image quality compared with FBP and HIR. • Full MBIR leads to image texture change which reduces the confidence in these images among radiologists. • Awareness of the image texture change and improved quality of full MBIR reconstructed images could improve the acceptance of this technique in clinical practice. [ABSTRACT FROM AUTHOR]

Background: Brain CT needs more attention to improve the extremely low image contrast and image texture. Purpose: To evaluate the performance of iterative progressive reconstruction with visual modeling (IPV) for the improvement of low-contrast detectability (IPV-LCD) compared with filtered backprojection (FBP) and conventional IPV. Materials and methods: Low-contrast and water phantoms were used. Helical scans were conducted with the use of a CT scanner with 64 detectors. The tube voltage was set at 120 kVp; the tube current was adjusted from 60 to 300 mA with a slice thickness of 0.625 mm and from 20 to 150 mA with a slice thickness of 5.0 mm. Images were reconstructed with the FBP, conventional IPV, and IPV-LCD algorithms. The channelized Hotelling observer (CHO) model was applied in conjunction with the use of low-contrast modules in the low-contrast phantom. The noise power spectrum (NPS) and normalized NPS were calculated. Results: At the same standard and strong levels, the IPV-LCD method improved low-contrast detectability compared with the conventional IPV, regardless of contrast-rod diameters. The mean CHO values at a slice thickness of 0.625 mm were 1.83, 3.28, 4.40, 4.53, and 5.27 for FBP, IPV STD, IPV-LCD STD, IPV STR, and IPV-LCD STR, respectively. The normalized NPS for the IPV-LCD STD and STR images were slightly shifted to the higher frequency compared with that for the FBP image. Conclusion: IPV-LCD images further improve the low-contrast detectability compared with FBP and conventional IPV images while maintaining similar FBP image appearances. [ABSTRACT FROM AUTHOR]

Purpose: Phantoms are a basic tool for assessing and verifying performance in CT research and clinical practice. Patient‐based realistic lung phantoms accurately representing textures and densities are essential in developing and evaluating novel CT hardware and software. This study introduces PixelPrint, a 3D printing solution to create patient‐based lung phantoms with accurate attenuation profiles and textures. Methods: PixelPrint, a software tool, was developed to convert patient digital imaging and communications in medicine (DICOM) images directly into FDM printer instructions (G‐code). Density was modeled as the ratio of filament to voxel volume to emulate attenuation profiles for each voxel, with the filament ratio controlled through continuous modification of the printing speed. A calibration phantom was designed to determine the mapping between filament line width and Hounsfield units (HU) within the range of human lungs. For evaluation of PixelPrint, a phantom based on a single human lung slice was manufactured and scanned with the same CT scanner and protocol used for the patient scan. Density and geometrical accuracy between phantom and patient CT data were evaluated for various anatomical features in the lung. Results: For the calibration phantom, measured mean HU show a very high level of linear correlation with respect to the utilized filament line widths, (r > 0.999). Qualitatively, the CT image of the patient‐based phantom closely resembles the original CT image both in texture and contrast levels (from −800 to 0 HU), with clearly visible vascular and parenchymal structures. Regions of interest comparing attenuation illustrated differences below 15 HU. Manual size measurements performed by an experienced thoracic radiologist reveal a high degree of geometrical correlation of details between identical patient and phantom features, with differences smaller than the intrinsic spatial resolution of the scans. Conclusion: The present study demonstrates the feasibility of 3D‐printed patient‐based lung phantoms with accurate organ geometry, image texture, and attenuation profiles. PixelPrint will enable applications in the research and development of CT technology, including further development in radiomics. [ABSTRACT FROM AUTHOR]

• ImmunoPET is a potential biomarker for patient selection to immunotherapy. • Patient response classification differs with tumor measurement strategy. • MPM tumor volumetry predicts patient survival and outperforms CT volumetry. • CT image texture has potential to differentiate among MPM tumor histologies. • Skeletal muscle loss in chemotherapy patients demonstrates prognostic potential. Imaging of mesothelioma plays a role in all aspects of patient management, including disease detection, staging, evaluation of treatment options, response assessment, pre-surgical evaluation, and surveillance. Imaging in this disease impacts a wide range of disciplines throughout the healthcare enterprise. Researchers and clinician-scientists are developing state-of-the-art techniques to extract more of the information contained within these medical images and to utilize it for more sophisticated tasks; moreover, image-acquisition technology is advancing the inherent capabilities of these images. This paper summarizes the imaging-based topics presented orally at the 2021 International Conference of the International Mesothelioma Interest Group (iMig), which was held virtually from May 7–9, 2021. These topics include an update on the mesothelioma staging system, novel molecular targets to guide therapy in mesothelioma, special considerations and potential pitfalls in imaging mesothelioma in the immunotherapy setting, tumor measurement strategies and their correlation with patient survival, tumor volume measurement in MRI and CT, CT-based texture analysis for differentiation of histologic subtype, diffusion-weighted MRI for the assessment of biphasic mesothelioma, and the prognostic significance of skeletal muscle loss with chemotherapy. [ABSTRACT FROM AUTHOR]

Purpose: In contrast-enhanced abdominopelvic CT (CE-APCT) for oncologic follow-up, ultrahigh-resolution CT (UHRCT) may improve depiction of fine lesions and low-dose scans are desirable for minimizing the potential adverse effects by ionizing radiation. We compared image quality and radiologists' acceptance of model-based iterative (MBIR) and deep learning (DLR) reconstructions of low-dose CE-APCT by UHRCT. Methods: Using our high-resolution (matrix size: 1024) and low-dose (tube voltage 100 kV; noise index: 20–40 HU) protocol, we scanned phantoms to compare the modulation transfer function and noise power spectrum between MBIR and DLR and assessed findings in 36 consecutive patients who underwent CE-APCT (noise index: 35 HU; mean CTDIvol: 4.2 ± 1.6 mGy) by UHRCT. We used paired t-test to compare objective noise and contrast-to-noise ratio (CNR) and Wilcoxon signed-rank test to compare radiologists' subjective acceptance regarding noise, image texture and appearance, and diagnostic confidence between MBIR and DLR using our routine protocol (matrix size: 512; tube voltage: 120 kV; noise index: 15 HU) for reference. Results: Phantom studies demonstrated higher spatial resolution and lower low-frequency noise by DLR than MBIR at equal doses. Clinical studies indicated significantly worse objective noise, CNR, and subjective noise by DLR than MBIR, but other subjective characteristics were better (P < 0.001 for all). Compared with the routine protocol, subjective noise was similar or better by DLR, and other subjective characteristics were similar or worse by MBIR. Conclusion: Image quality, except regarding noise characteristics, and acceptance by radiologists were better by DLR than MBIR in low-dose CE-APCT by UHRCT. [ABSTRACT FROM AUTHOR]

Local binary patterns (LBPs) are used for effective texture representation in various applications. This study explores the clustering consistency and stability of image segmentation when distance‐based clustering methods are used with an LBP. Because data are described by features and attributes, distances among data are also dominated by the definition of features. Moreover, four popular LBP encoding schemes for segmenting computed tomography (CT) images by using fuzzy C‐means are discussed and compared. The experimental results demonstrate several notable phenomena: When the pixels are encoded in different LBP encoding schemes, the distance between these pixels varies considerably. The experimental results indicate that each LBP encoding scheme emphasizes a specific image texture that dominates how pixels are described in the feature space and thus affects the clustering results. Based on the evaluation of CT image segmentation by using the clustering inconsistency index, linear and corner‐like LBPs are particularly suitable for clustering longitudinal sections and cross sections of CT images, respectively. [ABSTRACT FROM AUTHOR]

Purpose: The low‐dose computed tomography (CT) imaging can reduce the damage caused by x‐ray radiation to the human body. However, low‐dose CT images have a different degree of artifacts than conventional CT images, and their resolution is lower than that of conventional CT images, which can affect disease diagnosis by clinicians. Therefore, methods for noise‐level reduction and resolution improvement in low‐dose CT images have inevitably become a research hotspot in the field of low‐dose CT imaging. Methods: In this paper, residual attention modules (RAMs) are incorporated into the residual encoder–decoder convolutional neural network (RED‐CNN) and generative adversarial network with Wasserstein distance (WGAN) to learn features that are beneficial to improving the performances of denoising networks, and developed models are denoted as RED‐CNN‐RAM and WGAN‐RAM, respectively. In detail, RAM is composed of a multi‐scale convolution module and an attention module built on the residual network architecture, where the attention module consists of a channel attention module and a spatial attention module. The residual network architecture solves the problem of network degradation with increased network depth. The function of the attention module is to learn which features are beneficial to reduce the noise level of low‐dose CT images to reduce the loss of detail in the final denoising images, which is also the key point of the proposed algorithms. Results: To develop a robust network for low‐dose CT image denoising, multidose‐level torso phantom images provided by a cooperating equipment vendor are used to train the network, which can improve the network's adaptability to clinical application. In addition, a clinical dataset is used to test the network's migration capabilities and clinical applicability. The experimental results demonstrate that these proposed networks can effectively remove noise and artifacts from multidose CT scans. Subjective and objective analyses of multiple groups of comparison experiments show that the proposed networks achieve good noise suppression performance while preserving the image texture details. Conclusion: In this study, two deep learning network models are developed using multidose‐level CT images acquired from a commercial spiral CT scanner. The two network models can reduce and even remove streaking artifacts, and noise from low‐dose CT images confirms the effectiveness of the proposed algorithms. [ABSTRACT FROM AUTHOR]

Purpose: Whole-body positron emission tomography/computed tomography (PET/CT) has become a standard method of imaging patients with various disease conditions, especially cancer. Body-wide accurate quantification of disease burden in PET/CT images is important for characterizing lesions, staging disease, prognosticating patient outcome, planning treatment, and evaluating disease response to therapeutic interventions. However, body-wide anatomy recognition in PET/CT is a critical first step for accurately and automatically quantifying disease body-wide, body-region-wise, and organwise. This latter process, however, has remained a challenge due to the lower quality of the anatomic information portrayed in the CT component of this imaging modality and the paucity of anatomic details in the PET component. In this paper, the authors demonstrate the adaptation of a recently developed automatic anatomy recognition (AAR) methodology [Udupa et al., "Body-wide hierarchical fuzzy modeling, recognition, and delineation of anatomy in medical images," Med. Image Anal. 18, 752-771 (2014)] to PET/CT images. Their goal was to test what level of object localization accuracy can be achieved on PET/CT compared to that achieved on diagnostic CT images. Methods: The authors advance the AAR approach in this work in three fronts: (i) from body-regionwise treatment in the work of Udupa et al. to whole body; (ii) from the use of image intensity in optimal object recognition in the work of Udupa et al. to intensity plus object-specific texture properties, and (iii) from the intramodality model-building-recognition strategy to the intermodality approach. The whole-body approach allows consideration of relationships among objects in different body regions, which was previously not possible. Consideration of object texture allows generalizing the previous optimal threshold-based fuzzy model recognition method from intensity images to any derived fuzzy membership image, and in the process, to bring performance to the level achieved on diagnostic CT and MR images in body-region-wise approaches. The intermodality approach fosters the use of already existing fuzzy models, previously created from diagnostic CT images, on PET/CT and other derived images, thus truly separating the modality-independent object assembly anatomy from modality-specific tissue property portrayal in the image. Results: Key ways of combining the above three basic ideas lead them to 15 different strategies for recognizing objects in PET/CT images. Utilizing 50 diagnostic CT image data sets from the thoracic and abdominal body regions and 16 whole-body PET/CT image data sets, the authors compare the recognition performance among these 15 strategies on 18 objects from the thorax, abdomen, and pelvis in object localization error and size estimation error. Particularly on texture membership images, object localization is within three voxels on whole-body low-dose CT images and 2 voxels on body-region-wise low-dose images of known true locations. Surprisingly, even on direct body-region-wise PET images, localization error within 3 voxels seems possible. Conclusions: The previous body-region-wise approach can be extended to whole-body torso with similar object localization performance. Combined use of image texture and intensity property yields the best object localization accuracy. In both body-region-wise and whole-body approaches, recognition performance on low-dose CT images reaches levels previously achieved on diagnostic CT images. The best object recognition strategy varies among objects; the proposed framework however allows employing a strategy that is optimal for each object. [ABSTRACT FROM AUTHOR]

Purpose: Preoperative prediction of perineural invasion (PNI) and Kirsten RAS (KRAS) mutation in colon cancer is critical for treatment planning and patient management. We developed machine learning models for diagnosis of PNI and KRAS mutation in colon cancer patients by interpreting preoperative CT. Methods: This retrospective study included 207 patients who received surgical resection in our institution. The underlying tumor characteristics were described by analyzing CT image texture quantitatively. The key radiomics features were determined with similarity analysis followed by RELIEFF method among 306 CT imaging features. Eight kernel-based support vector machines classifiers were constructed using individual (II, III, or IV) or multi-stage (II + III + IV) patient cohorts for predicting PNI and KRAS mutation. The model performances were evaluated using accuracy, receiver operating curve, and decision curve analyses. Results: Multi-stage classifiers obtained AUC of 0.793 and 0.862 for detecting PNI and KRAS mutation for test cohort. Moreover, individual-stage classifiers demonstrated significantly improved diagnostic performance at all stages (IIAUC: [0.86; 0.99], IIIAUC: [0.99; 0.99], and IVAUC: [1.00; 1.00], respectively, for PNI and KRAS mutation in test cohort). Besides, stage II tumor is better described with coarse texture features while more detailed features are required for better characterization of advanced-stage tumors (III and IV) for diagnoses of PNI or KRAS mutation. Conclusion: Machine learning models developed using preoperative CT data can predict PNI and KRAS mutation in colon cancer patients with satisfactory performance. Individual-stage models better-characterized the relationship between CT features and PNI or KRAS mutation than multi-stage models and demonstrated good prediction scores. [ABSTRACT FROM AUTHOR]

To investigate whether the FD of non-small cell lung cancer (NSCLC) on CT predicts tumor stage and uptake on 18F-fluorodeoxyglucose positron emission tomography. The FD within a tumor region was determined using a box counting algorithm and compared to the lymph node involvement (NI) and metastatic involvement (MI) and overall stage as determined from PET. A Mann-Whitney U test was applied to the extracted FD features for the NI and the MI. The two tests showed good significance with p <.05 (p NI = 0.0139, p MI = 0.0194). Also after performing fractal analysis to all cases, it was found that for those who had a CT of stage I or II had a higher likelihood of the NI and/or MI stage being upstaged by PET, Odds Ratio 5.38 (95% CI 0.99–29.3). For those who are CT stage III or IV had an increased likelihood of the NI and/or MI stage being down staged by PET, Odds Ratio: 7.33 (95% CI 0.48–111.2). Initial results of this study indicate higher FD in CT images of NSCLC is associated with advanced stage and greater FDG uptake on PET. Measurements of tumor fractal analysis on conventional non-contrast CT examinations could potentially be used as a prognostic marker and/or to select patients for PET. Unlabelled Image • Fractal analysis can quantify complexity within image texture. • Conventional non-contrast enhanced CT scans are employed. • Fractal dimension (FD) estimated within a tumor region using box counting algorithm • FD values correlate with PET lymph node, metastasis and overall tumor stage. • Fractal analysis could potentially be used as a prognostic marker and/or to select patients for PET. [ABSTRACT FROM AUTHOR]

Purpose: Benign computed tomography (CT) changes due to radiation induced lung injury (RILI) are common following stereotactic ablative radiotherapy (SABR) and can be difficult to differentiate from tumor recurrence. The authors measured the ability of CT image texture analysis, compared to more traditional measures of response, to predict eventual cancer recurrence based on CT images acquired within 5 months of treatment. Methods: A total of 24 lesions from 22 patients treated with SABR were selected for this study: 13 with moderate to severe benign RILI, and 11 with recurrence. Three-dimensional (3D) consolidative and ground-glass opacity (GGO) changes were manually delineated on all follow-up CT scans. Two size measures of the consolidation regions (longest axial diameter and 3D volume) and nine appearance features of the GGO were calculated: 2 first-order features [mean density and standard deviation of density (first-order texture)], and 7 second-order texture features [energy, entropy, correlation, inverse difference moment (IDM), inertia, cluster shade, and cluster prominence]. For comparison, the corresponding response evaluation criteria in solid tumors measures were also taken for the consolidation regions. Prediction accuracy was determined using the area under the receiver operating characteristic curve (AUC) and two-fold cross validation (CV). Results: For this analysis, 46 diagnostic CT scans scheduled for approximately 3 and 6 months posttreatment were binned based on their recorded scan dates into 2-5 month and 5-8 month follow-up time ranges. At 2-5 months post-treatment, first-order texture, energy, and entropy provided AUCs of 0.79-0.81 using a linear classifier. On two-fold CV, first-order texture yielded 73% accuracy versus 76%-77% with the second-order features. The size measures of the consolidative region, longest axial diameter and 3D volume, gave two-fold CV accuracies of 60% and 57%, and AUCs of 0.72 and 0.65, respectively. Conclusions: Texture measures of the GGO appearance following SABR demonstrated the ability to predict recurrence in individual patients within 5 months of SABR treatment. Appearance changes were also shown to be more accurately predictive of recurrence, as compared to size measures within the same time period. With further validation, these results could form the substrate for a clinically useful computer-aided diagnosis tool which could provide earlier salvage of patients with recurrence. [ABSTRACT FROM AUTHOR]

This paper studies the sensitivity of a range of image texture parameters used in radiomics to: (i) the number of intensity levels, (ii) the method of quantisation to select the intensity levels and (iii) the use of an intensity threshold. 43 commonly used texture features were studied for the gross target volume outlined on the CT component of PET/CT scans of 50 patients with non-small cell lung carcinoma (NSCLC). All cases were quantised for all values between 4 and 128 intensity levels using four commonly used quantisation methods. All results were analysed with and without a threshold range of −200 HU to 300 HU. Cases were ranked for each texture feature and for all quantisation methods with the Spearman's rank correlation coefficient determined to evaluate stability. Results showed large fluctuations in ranking, particularly for low numbers of levels, differences between quantisation methods and with the use of a threshold, with values Spearman's Rank Correlation for many parameters below 0.2. Our results demonstrated the sensitivity of radiomics features to the parameters used during analysis and highlight the risk of low reproducibility comparing studies with slightly different parameters. In terms of the lung cancer CT datasets, this study supports the use of 128 intensity levels, the same uniform quantiser applied to all scans and thresholding of the data. It also supports several of the features recommended in the literature for such studies such as skewness and kurtosis. A recommended framework is presented for curation of the data analysis process to ensure stability of results. [ABSTRACT FROM AUTHOR]

Objectives: To determine whether image texture parameters analysed on pre-operative contrast-enhanced computed tomography (CT) can predict overall survival and recurrence-free survival in patients with hepatocellular carcinoma (HCC) treated by surgical resection.Methods: We retrospectively included all patients operated for HCC who had liver contrast-enhanced CT within 3 months prior to treatment in our centre between 2010 and 2015. The following texture parameters were evaluated on late-arterial and portal-venous phases: mean grey-level, standard deviation, kurtosis, skewness and entropy. Measurements were made before and after spatial filtration at different anatomical scales (SSF) ranging from 2 (fine texture) to 6 (coarse texture). Lasso penalised Cox regression analyses were performed to identify independent predictors of overall survival and recurrence-free survival.Results: Forty-seven patients were included. Median follow-up time was 345 days (interquartile range [IQR], 176-569). Nineteen patients had a recurrence at a median time of 190 days (IQR, 141-274) and 13 died at a median time of 274 days (IQR, 96-411). At arterial CT phase, kurtosis at SSF = 4 (hazard ratio [95% confidence interval] = 3.23 [1.35-7.71] p = 0.0084) was independent predictor of overall survival. At portal-venous phase, skewness without filtration (HR [CI 95%] = 353.44 [1.31-95102.23], p = 0.039), at SSF2 scale (HR [CI 95%] = 438.73 [2.44-78968.25], p = 0.022) and SSF3 (HR [CI 95%] = 14.43 [1.38-150.51], p = 0.026) were independently associated with overall survival. No textural feature was identified as predictor of recurrence-free survival.Conclusions: In patients with resectable HCC, portal venous phase-derived CT skewness is significantly associated with overall survival and may potentially become a useful tool to select the best candidates for resection.Key Points: • HCC heterogeneity as evaluated by texture analysis of contrast-enhanced CT images may predict overall survival in patients treated by surgical resection. • Among texture parameters, skewness assessed at different anatomical scales at portal-venous phase CT is an independent predictor of overall survival after resection. • In patients with HCC, CT texture analysis may have the potential to become a useful tool to select the best candidates for resection. [ABSTRACT FROM AUTHOR]

• DLR images achieve substantial noise reduction compared with other reconstruction algorithms. • DLR images maintain noise texture better than other reconstruction algorithms. • Sole phantom studies are insufficient for future evaluation of DLR techniques. To compare the noise power spectrum (NPS) properties and perform a qualitative analysis of hybrid iterative reconstruction (IR), model-based IR (MBIR), and deep learning–based reconstruction (DLR) at a similar noise level in clinical study and compare these outcomes with those in phantom study. A Catphan phantom with an external body ring was used in the phantom study. In the clinical study, computed tomography (CT) examination data of 34 patients were reviewed. NPS was calculated from DLR, hybrid IR, and MBIR images. The noise magnitude ratio (NMR) and the central frequency ratio (CFR) were calculated from DLR, hybrid IR, and MBIR images relative to filtered back-projection images using NPS. Clinical images were independently reviewed by two radiologists. In the phantom study, DLR with a mild level had a similar noise level as hybrid IR and MBIR with strong levels. In the clinical study, DLR with a mild level had a similar noise level as hybrid IR with standard and MBIR with strong levels. The NMR and CFR were 0.40 and 0.76 for DLR, 0.42 and 0.55 for hybrid IR, and 0.48 and 0.62 for MBIR. The visual inspection of the clinical DLR image was superior to that of the hybrid IR and MBIR images. Deep learning–based reconstruction improves overall image quality with substantial noise reduction while maintaining image noise texture compared with the CT reconstruction techniques. [ABSTRACT FROM AUTHOR]

Objective: To objectively quantify intracranial hematoma (ICH) enlargement by analysing the image texture of head CT scans and to provide objective and quantitative imaging parameters for predicting early hematoma enlargement.Methods: We retrospectively studied 108 ICH patients with baseline non-contrast computed tomography (NCCT) and 24-h follow-up CT available. Image data were assessed by a chief radiologist and a resident radiologist. Consistency analysis between observers was tested. The patients were divided into training set (75%) and validation set (25%) by stratified sampling. Patients in the training set were dichotomized according to 24-h hematoma expansion ≥ 33%. Using the Laplacian of Gaussian bandpass filter, we chose different anatomical spatial domains ranging from fine texture to coarse texture to obtain a series of derived parameters (mean grayscale intensity, variance, uniformity) in order to quantify and evaluate all data. The parameters were externally validated on validation set.Results: Significant differences were found between the two groups of patients within variance at V1.0 and in uniformity at U1.0, U1.8 and U2.5. The intraclass correlation coefficients for the texture parameters were between 0.67 and 0.99. The area under the ROC curve between the two groups of ICH cases was between 0.77 and 0.92. The accuracy of validation set by CTTA was 0.59-0.85.Conclusion: NCCT texture analysis can objectively quantify the heterogeneity of ICH and independently predict early hematoma enlargement.Key Points: • Heterogeneity is helpful in predicting ICH enlargement. • CTTA could play an important role in predicting early ICH enlargement. • After filtering, fine texture had the best diagnostic performance. • The histogram-based uniformity parameters can independently predict ICH enlargement. • CTTA is more objective, more comprehensive, more independently operable, than previous methods. [ABSTRACT FROM AUTHOR]

Purpose: To compare the predictive models that can incorporate a set of CT image features for preoperatively differentiating the high grade (Fuhrman III-IV) from low grade (Fuhrman I-II) clear cell renal cell carcinoma (ccRCC).Material and Methods: One hundred and fourteen patients with ccRCC treated with a partial or radical nephrectomy were enrolled in the training cohort. The six non-texture features, including Pseudocapsule, Round mass, maximal tumor diameter (Diametermax), intratumoral artery (Arterytumor), enhancement value of the tumor (TEV) and relative TEV (rTEV), were assessed for each tumor. The texture features were extracted from the CT images of the section with the largest area of renal mass at both corticomedullary and nephrographic phases. The least absolute shrinkage and selection operator (LASSO) was used to screen the most valuable texture features to calculate a texture score (Texture-score) for each patient. A logistic regression model was used in the training cohort to discriminate the high from low grade ccRCC at nephrectomy. The predictors would include all non-texture features in Model 1, all non-texture features and Texture-score in Model 2, and Texture-score in Model 3. The performance of the predictive models were tested and compared in an independent validation cohort composed of 92 cases with ccRCC.Results: Inter-rater agreement was good for each non-texture feature and Texture-score (the concordance correlation coefficient or Kappa coefficient > 0.70). The Texture-score was calculated via a linear combination of the 4 selected texture features. The three models shown good discrimination of the high from low grade ccRCC in the training cohort and the area under receiver operating characteristic curve (AUC) was 0.826 in Mode 1, 0.878 in Model 2 and 0.843 in Model 3, and a significant different AUC was found between Model 1 and Model 2. Application of the predictive models in the validation cohort still gave a discrimination (AUC > 0.670), and the Texture-score based models with or without the non-texture features (Model 2 and 3) showed a better discrimination of the high from low grade ccRCC (P < 0.05).Conclusion: This study presented the Texture-score based models can facilitate the preoperative discrimination of the high from low grade ccRCC. [ABSTRACT FROM AUTHOR]

Osteoporosis is characterized by bone loss and degradation of bone microstructure leading to fracture particularly in elderly people. Osteoporotic bone degeneration and fracture risk can be assessed by bone mineral density and trabecular bone score from 2D projection dual-energy X-ray absorptiometry images. However, multidetector computed tomography image based quantification of trabecular bone microstructure showed significant improvement in prediction of fracture risk beyond that from bone mineral density and trabecular bone score; however, high radiation exposure limits its use in routine clinical in vivo examinations. Hence, this study investigated reduction of radiation dose and its effects on image quality of thoracic midvertebral specimens. Twenty-four texture features were extracted to quantify the image quality from multidetector computed tomography images of 11 thoracic midvertebral specimens, by means of statistical moments, the gray-level co-occurrence matrix, and the gray-level run-length matrix, and were analyzed by an independent sample t-test to observe differences in image texture with respect to radiation doses of 80, 150, 220, and 500 mAs. The results showed that three features-namely, global variance, energy, and run percentage, were not statistically significant ([Formula: see text]) for low doses with respect to 500 mAs. Hence, it is evident that these three dose-independent features can be used for disease monitoring with a low-dose imaging protocol. [ABSTRACT FROM AUTHOR]

Simple Summary: Pediatric osteosarcoma is one of the most aggressive cancers, and predictions of metastasis and chemotherapy response have a significant impact on pediatric patient survival. Radiogenomics, as methods of analyzing gene expression or image texture features, have previously been used for the diagnosis of chemotherapy responses and metastasis and can reveal the current state of cancer. In this study, we aimed to generate a predictive model using gene expression and 18F-FDG PET/CT image texture features in pediatric osteosarcoma in relation to metastasis and chemotherapy response. A predictive model using radiogenomics technology that incorporates both imaging features and gene expression can accurately predict metastasis and chemotherapy responses to improve patient outcomes. Chemotherapy response and metastasis prediction play important roles in the treatment of pediatric osteosarcoma, which is prone to metastasis and has a high mortality rate. This study aimed to estimate the prediction model using gene expression and image texture features. 18F-fluorodeoxyglucose positron emission tomography/computed tomography (18F-FDG PET/CT) images of 52 pediatric osteosarcoma patients were used to estimate the machine learning algorithm. An appropriate algorithm was selected by estimating the machine learning accuracy. 18F-FDG PET/CT images of 21 patients were selected for prediction model development based on simultaneous KI67 and EZRIN expression. The prediction model for chemotherapy response and metastasis was estimated using area under the curve (AUC) maximum image texture features (AUC_max) and gene expression. The machine learning algorithm with the highest test accuracy in chemotherapy response and metastasis was selected using the random forest algorithm. The chemotherapy response and metastasis test accuracy with image texture features was 0.83 and 0.76, respectively. The highest test accuracy and AUC of chemotherapy response with AUC_max, KI67, and EZRIN were estimated to be 0.85 and 0.89, respectively. The highest test accuracy and AUC of metastasis with AUC_max, KI67, and EZRIN were estimated to be 0.85 and 0.8, respectively. The metastasis prediction accuracy increased by 10% using radiogenomics data. [ABSTRACT FROM AUTHOR]

To balance noise reduction and texture details of low-dose computed tomography (LDCT) images during deep learning training, we propose a network combining edge features and a pyramid structure for LDCT denoising (PLEDNet). PLEDNet contains three states: superficial feature acquisition, deep feature acquisition, and feature confluence. First, the superficial feature acquisition stage uses a simple convolutional block to perform feature extraction on the input LDCT image. Second, in the deep feature extraction phase, one branch uses pyramid pooling to extract multiscale features, and the other uses the trainable Laplacian of the Gaussian module to extract edge detail features. Finally, the feature fusion phase uses a fusion attention module to adaptively fuse multiscale and edge features. The results of the experiments demonstrate that the method not only enhances various relative objective indicators but also effectively represses the noise in LDCT images while fully conserving the image texture details, which has great advantages in terms of visual effects. [ABSTRACT FROM AUTHOR]

Background: Image texture analysis is a noninvasive technique for quantifying intratumoral heterogeneity, with derived texture features reported to be closely related to the treatment outcome of tumors. Gastric cancer is one of the most common tumors and the third leading cause of cancer-related deaths worldwide. Although trastuzumab is associated with a survival gain among patients with human epidermal growth factor receptor 2 (HER2)-positive advanced gastric cancer, optimal patient selection is challenging. The purpose of this study was to determine whether CT texture features of HER2-positive gastric cancer were related to the survival rate after trastuzumab treatment. Methods and Findings: Patients diagnosed with HER2-positive advanced gastric cancer from February 2007 to August 2014 were retrospectively selected. Using in-house built software, histogram features (kurtosis and skewness) and gray-level co-occurrence matrices (GLCM) features (angular second moment [ASM], contrast, entropy, variance, and correlation) were derived from the CT images of HER2-positive advanced gastric cancer in 26 patients. All the patients were followed up for more than 6 months, with no confirmed deaths. The patients were dichotomized into a good and poor survival group based on cutoff points of overall survival of 12 months. A receiver-operating characteristics (ROC) analysis was performed to test the ability of each texture parameter to identify the good survival group. Kaplan–Meier curves for patients above and below each threshold were constructed. Using a threshold of >265.8480 for contrast, >488.3150 for variance, and ≤0.1319×10−3. for correlation, all of the area under the ROC curves showed fair accuracy (>0.7). Kaplan–Meier analysis showed statistically significant survival difference between two groups according to optimal cutoff values of contrast, variance, correlation and ASM. However, as this study had a small number of patients, a further study with a larger population will be needed to validate the results. Conclusions: Heterogeneous texture features on CT images were associated with better survival in patients with HER2-positive advanced gastric cancer who received trastuzumab-based treatment. Therefore, texture analysis shows potential to be a clinically useful imaging biomarker providing additional prognostic information for patient selection. [ABSTRACT FROM AUTHOR]

Introduction: The purpose of this study was to evaluate the utility of iterative model reconstruction (IMR) in brain CT especially with thin-slice images. Methods: This prospective study received institutional review board approval, and prior informed consent to participate was obtained from all patients. We enrolled 34 patients who underwent brain CT and reconstructed axial images with filtered back projection (FBP), hybrid iterative reconstruction (HIR) and IMR with 1 and 5 mm slice thicknesses. The CT number, image noise, contrast, and contrast noise ratio (CNR) between the thalamus and internal capsule, and the rate of increase of image noise in 1 and 5 mm thickness images between the reconstruction methods, were assessed. Two independent radiologists assessed image contrast, image noise, image sharpness, and overall image quality on a 4-point scale. Results: The CNRs in 1 and 5 mm slice thickness were significantly higher with IMR (1.2 ± 0.6 and 2.2 ± 0.8, respectively) than with FBP (0.4 ± 0.3 and 1.0 ± 0.4, respectively) and HIR (0.5 ± 0.3 and 1.2 ± 0.4, respectively) ( p < 0.01). The mean rate of increasing noise from 5 to 1 mm thickness images was significantly lower with IMR (1.7 ± 0.3) than with FBP (2.3 ± 0.3) and HIR (2.3 ± 0.4) ( p < 0.01). There were no significant differences in qualitative analysis of unfamiliar image texture between the reconstruction techniques. Conclusion: IMR offers significant noise reduction and higher contrast and CNR in brain CT, especially for thin-slice images, when compared to FBP and HIR. [ABSTRACT FROM AUTHOR]

Background Food microstructure can be visualised by a wide range of microscopic techniques, however these methods are usually destructive and require sample preparation. X-ray micro-computed tomography (μCT) provides an alternative as it is non-invasive, non-destructive and requires no sample preparation. It characterises structures three-dimensionally, allowing evaluation of microstructural changes at resolutions as high as a few hundred nanometres. After the discovery of X-rays in 1895, X-ray computed tomography (CT) was developed and introduced into clinical practices in the 1970s. The first X-ray μCT food application, to detect the maturity of green tomatoes, followed in 1991. Scope and approach This review aims to provide an overview of the basic principles of X-ray μCT, the different systems, image processing and analysis as well as image texture analysis. Food applications are highlighted and the review concludes with future trends of X-ray μCT. Key findings and conclusions The controlled production and stability of microstructure is of great interest to the food industry. Both laboratory μCT and synchrotron systems are becoming more common and thus will lead to imaging in three dimensions at a micron scale playing a much bigger role in future food studies. Limitations include operator dependency, time and cost constraints and imaging artefacts. Technological and computational progress, however, encourages the growth of this technique in food science. [ABSTRACT FROM AUTHOR]

Introduction This study aims to explore whether the intra-tumour 18F-fluorodeoxyglucose ( FDG) uptake heterogeneity affects the reliability of target volume definition with FDG positron emission tomography/computed tomography ( PET/ CT) imaging for nonsmall cell lung cancer ( NSCLC) and squamous cell oesophageal cancer ( SCEC). Methods Patients with NSCLC ( n = 50) or SCEC ( n = 50) who received 18F- FDG PET/ CT scanning before treatments were included in this retrospective study. Intra-tumour FDG uptake heterogeneity was assessed by visual scoring, the coefficient of variation ( COV) of the standardised uptake value ( SUV) and the image texture feature (entropy). Tumour volumes (gross tumour volume ( GTV) ) were delineated on the CT images ( GTVCT), the fused PET/ CT images ( GTVPET-CT) and the PET images, using a threshold at 40% SUVmax ( GTVPET40%) or the SUV cut-off value of 2.5 ( GTVPET2.5). The correlation between the FDG uptake heterogeneity parameters and the differences in tumour volumes among GTVCT, GTVPET-CT, GTVPET40% and GTVPET2.5 was analysed. Results For both NSCLC and SCEC, obvious correlations were found between uptake heterogeneity, SUV or tumour volumes. Three types of heterogeneity parameters were consistent and closely related to each other. Substantial differences between the four methods of GTV definition were found. The differences between the GTV correlated significantly with PET heterogeneity defined with the visual score, the COV or the textural feature-entropy for NSCLC and SCEC. Conclusions In tumours with a high FDG uptake heterogeneity, a larger GTV delineation difference was found. Advance image segmentation algorithms dealing with tracer uptake heterogeneity should be incorporated into the treatment planning system. [ABSTRACT FROM AUTHOR]

To assess the changes in computed tomography ( CT) tumor heterogeneity following neoadjuvant chemotherapy in esophageal cancer. Thirty-one consecutive patients who received neoadjuvant chemotherapy for esophageal cancer were identified. Analysis of primary tumor heterogeneity (texture) was performed on staging and post-chemotherapy CT scans. Image texture parameters (mean grey-level intensity, entropy, uniformity, kurtosis, skewness, standard deviation of histogram) were derived for different levels of image filtration (0-2.5). Proportional changes in each parameter following treatment were obtained. Comparison between pathological tumor response and texture parameters was analyzed using Mann- Whitney U-test. The relationship between CT texture and overall survival) was estimated using the Kaplan- Meier method. Tumor texture became more homogeneous after treatment with a significant decrease in entropy and increase in uniformity (filter 1.0 and 2.5). Pretreatment (filter 1.5, P = 0.006) and posttreatment standard deviation of histogram (filter 1.0, P = 0.009) showed a borderline association with pathological tumor response. A proportional change in skewness <0.39 (filter 1.0) was associated with improved survival (median overall survival 36.1 vs. 11.1 months; P < 0.001). CT tumor heterogeneity decreased following neoadjuvant chemotherapy and has the potential to provide additional information in primary esophageal cancer. [ABSTRACT FROM AUTHOR]

Abstract: An abdominal wall hernia is a protrusion of the intestine through an opening or area of weakness in the abdominal wall. Correct pre-operative identification of abdominal wall hernia meshes could help surgeons adjust the surgical plan to meet the expected difficulty and morbidity of operating through or removing the previous mesh. First, we present herein for the first time the application of image analysis for automated identification of hernia meshes. Second, we discuss the novel development of a new entropy-based image texture feature using geostatistics and indicator kriging. Third, we seek to enhance the hernia mesh identification by combining the new texture feature with the gray-level co-occurrence matrix feature of the image. The two features can characterize complementary information of anatomic details of the abdominal hernia wall and its mesh on computed tomography. Experimental results have demonstrated the effectiveness of the proposed study. The new computational tool has potential for personalized mesh identification which can assist surgeons in the diagnosis and repair of complex abdominal wall hernias. [Copyright &y& Elsevier]

Limited-view Computed Tomography (CT) can be used to efficaciously reduce radiation dose in clinical diagnosis, it is also adopted when encountering inevitable mechanical and physical limitation in industrial inspection. Nevertheless, limited-view CT leads to severe artifacts in its imaging, which turns out to be a major issue in the low dose protocol. Thus, how to exploit the limited prior information to obtain high-quality CT images becomes a crucial issue. We notice that almost all existing methods solely focus on a single CT image while neglecting the solid fact that, the scanned objects are always highly spatially correlated. Consequently, there lies bountiful spatial information between these acquired consecutive CT images, which is still largely left to be exploited. In this paper, we propose a novel hybrid-domain structure composed of fully convolutional networks that groundbreakingly explores the three-dimensional neighborhood and works in a "coarse-to-fine" manner. We first conduct data completion in the Radon domain, and transform the obtained full-view Radon data into images through FBP. Subsequently, we employ the spatial correlation between continuous CT images to productively restore them and then refine the image texture to finally receive the ideal high-quality CT images, achieving PSNR of 40.209 and SSIM of 0.943. Besides, unlike other current limited-view CT reconstruction methods, we adopt FBP (and implement it on GPUs) instead of SART-TV to significantly accelerate the overall procedure and realize it in an end-to-end manner. [ABSTRACT FROM AUTHOR]

Total variation (TV) methods have been proposed to improve the image quality in count-reduced images, by reducing the variation between neighboring pixels. Although very easy to implement and fast to compute, TV-based methods may lead to a loss of texture information when applied to images with complex textures, such as high-resolution abdominal CT images. Here, we investigate the use of another regularization approach in the context of medical images based on multiresolution transformations. One such transformation is the shearlet transform, which is optimally sparse for images that are C^2 except for discontinuities along C^2 curves, and has better directional sensitivity than most other, related, wavelet transform approaches. We propose to solve the convex problem using the split-Bregman (augmented Lagrangian) approach. One of the primary advantages of the split-Bregman approach, is that the shearlet transform can easily be incorporated into the sparse-view CT reconstruction. The required sparsity prior is the \ell 1 norm of the shearlet coefficients. Results are shown for this method in comparison to the same framework with TV as the regularization term on simulated data. The noise-resolution performance is investigated at different contrast levels. At equal image noise, TV-based regularization outperforms shearlet-based regularization. However, when image texture is analyzed on measured mouse data, shearlets outperform TV, which suffers from staircasing effects. Our results show that there are benefits in using shearlets in CT imaging: texture is reconstructed more accurately compared to when TV is used, without biasing the image towards a piecewise constant image model. However, due to the larger support of the basis functions, our results suggest that uncareful usage of shearlets may lead to wavy artifacts, which can be equally unwanted as staircasing effects. [ABSTRACT FROM AUTHOR]

Background: Histology sections provide accurate information on atherosclerotic plaque composition, and are used in various applications. To our knowledge, no automated systems for plaque component segmentation in histology sections currently exist. Materials and Methods: We perform pixel wise classification of fibrous, lipid, and necrotic tissue in Elastica Von Gieson stained histology sections, using features based on color channel intensity and local image texture and structure. We compare an approach where we train on independent data to an approach where we train on one or two sections per specimen in order to segment the remaining sections. We evaluate the results on segmentation accuracy in histology, and we use the obtained histology segmentations to train plaque component classification methods in ex vivo Magnetic resonance imaging (MRI) and in vivo MRI and computed tomography (CT). Results: In leave one specimen out experiments on 176 histology slices of 13 plaques, a pixel wise accuracy of 75.7 ± 6.8% was obtained. This increased to 77.6 ± 6.5% when two manually annotated slices of the specimen to be segmented were used for training. Rank correlations of relative component volumes with manually annotated volumes were high in this situation (P = 0.82 0.98). Using the obtained histology segmentations to train plaque component classification methods in ex vivo MRI and in vivo MRI and CT resulted in similar image segmentations for training on the automated histology segmentations as for training on a fully manual ground truth. The size of the lipid rich necrotic core was significantly smaller when training on fully automated histology segmentations than when manually annotated histology sections were used. This difference was reduced and not statistically significant when one or two slices per section were manually annotated for histology segmentation. Conclusions: Good histology segmentations can be obtained by automated segmentation, which show good correlations with ground truth volumes. In addition, these can be used to develop segmentation methods in other imaging modalities. Accuracy increases when one or two sections of the same specimen are used for training, which requires a limited amount of user interaction in practice. [ABSTRACT FROM AUTHOR]

Background: The potential effects of ionizing radiation are of particular concern in children. The model-based iterative reconstruction VEO is a technique commercialized to improve image quality and reduce noise compared with the filtered back-projection (FBP) method. Objective: To evaluate the potential of VEO on diagnostic image quality and dose reduction in pediatric chest CT examinations. Materials and methods: Twenty children (mean 11.4 years) with cystic fibrosis underwent either a standard CT or a moderately reduced-dose CT plus a minimum-dose CT performed at 100 kVp. Reduced-dose CT examinations consisted of two consecutive acquisitions: one moderately reduced-dose CT with increased noise index (NI = 70) and one minimum-dose CT at CTDI 0.14 mGy. Standard CTs were reconstructed using the FBP method while low-dose CTs were reconstructed using FBP and VEO. Two senior radiologists evaluated diagnostic image quality independently by scoring anatomical structures using a four-point scale (1 = excellent, 2 = clear, 3 = diminished, 4 = non-diagnostic). Standard deviation (SD) and signal-to-noise ratio (SNR) were also computed. Results: At moderately reduced doses, VEO images had significantly lower SD ( P < 0.001) and higher SNR ( P < 0.05) in comparison to filtered back-projection images. Further improvements were obtained at minimum-dose CT. The best diagnostic image quality was obtained with VEO at minimum-dose CT for the small structures (subpleural vessels and lung fissures) ( P < 0.001). The potential for dose reduction was dependent on the diagnostic task because of the modification of the image texture produced by this reconstruction. Conclusions: At minimum-dose CT, VEO enables important dose reduction depending on the clinical indication and makes visible certain small structures that were not perceptible with filtered back-projection. [ABSTRACT FROM AUTHOR]

In recent years, interest in bone mineral densitometry (BMD) for children has increased, mainly due to the varieties of disease that influence bone growth. Trabecular bone densitometry in CT from lumbar vertebra is usually used clinically as a good indicator of bone deterioration and fracture risk. Nevertheless, diagnosing osteoporosis in obese and Duchenne Muscle Distrophy (DMD) children by BMD has been difficult due to the attenuation of radiation dose by abdominal fat, which could lead to overestimation of the degree of bone deterioration. The present work is an effort to improve pediatric osteoporosis diagnostic efficacy by analyzing the trabecular bone texture via CT with two frequency based methods: fractal dimension with power spectrum (Dps) and wavelet packets (WP). Healthy young adults group in the peak bone mass was statistically compared (t-student) with osteoporotic women with either fractal dimension or the four WP energy bands, and the results suggested significant differences. Applying ANOVA to three DMD children groups classified by their z-score as having normal, low and very low bone mineral density, significant differences were also found. At last, when comparing the DMD pediatric groups with that osteoporotic women, great statistical significance was found for all texture indicators. Results shown that the introduced techniques for image texture analysis could quantify the trabecular bone deterioration; this may help to improve the osteoporosis diagnosis in overweight patients, in specific DMD children, where it is often a diagnostic challenge. [ABSTRACT FROM AUTHOR]

• Our method takes into account the VOIs of the TCs, without the need to divide them into slices, i.e, we analyze the entire volume. • We propose eight image texture descriptors, which do not require parameterization and not need to resize images. • We have developed a scalable method, since it can be easily used in 2D or 3D images, without restrictions regarding the quantization of images. • Our descriptors can achieve results, as promising as deep networks, in some cases, with superior results. • Our descriptors, too, do not need large amounts of images to achieve good results. COVID-19 is an infectious disease caused by a newly discovered type of coronavirus called SARS-CoV-2. Since the discovery of this disease in late 2019, COVID-19 has become a worldwide concern, mainly due to its high degree of contagion. As of April 2021, the number of confirmed cases of COVID-19 reported to the World Health Organization has already exceeded 135 million worldwide, while the number of deaths exceeds 2.9 million. Due to the impacts of the disease, efforts in the literature have intensified in terms of studying approaches aiming to detect COVID-19, with a focus on supporting and facilitating the process of disease diagnosis. This work proposes the application of texture descriptors based on phylogenetic relationships between species to characterize segmented CT volumes, and the subsequent classification of regions into COVID-19, solid lesion or healthy tissue. To evaluate our method, we use images from three different datasets. The results are promising, with an accuracy of 99.93%, a recall of 99.93%, a precision of 99.93%, an F1-score of 99.93%, and an AUC of 0.997. We present a robust, simple, and efficient method that can be easily applied to 2D and/or 3D images without limitations on their dimensionality. [ABSTRACT FROM AUTHOR]

[Display omitted] • Proposed ImHistNet is capable of learning subtle textural features from raw images. • ImHistNet stratifies tumor and background textures, so no segmentation is required. • ImHistNet performs automatic RCC grade and stage classification. • We are the first, showing correlation between RCC stages and the deep-learned CT textural features. Fuhrman cancer grading and tumor-node-metastasis (TNM) cancer staging systems are typically used by clinicians in the treatment planning of renal cell carcinoma (RCC), a common cancer in men and women worldwide. Pathologists typically use percutaneous renal biopsy for RCC grading, while staging is performed by volumetric medical image analysis before renal surgery. Recent studies suggest that clinicians can effectively perform these classification tasks non-invasively by analyzing image texture features of RCC from computed tomography (CT) data. However, image feature identification for RCC grading and staging often relies on laborious manual processes, which is error prone and time-intensive. To address this challenge, this paper proposes a learnable image histogram in the deep neural network framework that can learn task-specific image histograms with variable bin centers and widths. The proposed approach enables learning statistical context features from raw medical data, which cannot be performed by a conventional convolutional neural network (CNN). The linear basis function of our learnable image histogram is piece-wise differentiable, enabling back-propagating errors to update the variable bin centers and widths during training. This novel approach can segregate the CT textures of an RCC in different intensity spectra, which enables efficient Fuhrman low (I/II) and high (III/IV) grading as well as RCC low (I/II) and high (III/IV) staging. The proposed method is validated on a clinical CT dataset of 159 patients from The Cancer Imaging Archive (TCIA) database, and it demonstrates 80% and 83% accuracy in RCC grading and staging, respectively. [ABSTRACT FROM AUTHOR]